Development of caregiver resilience scale (CRS) for Thai caregivers of older persons with dementia

Abstract This research aim to develop the Caregiver Resilience Scale (CRS) and examine the validity and reliability. The process began with a review of the concept of resilience based on a synthesis of existing research together with an exploration of qualitative data derived from an interview of ten caregivers of older persons with dementia. The CRS was examined by a panel of three experts to confirm its content validity, with the content validity index equal to 0.84. It was also tried out with 30 caregivers to test its internal consistency, and the result showed that the internal consistency was at a high level (Cronbach’s alpha 0.87). Furthermore, the Exploratory Factor Analysis was conducted with 150 caregivers to test construct validity of the scale, and it was found that all six domains of the CRS had construct validity. The final version of the CRS was composed of 30 items within six domains: physical competence; relationship competence; emotional competence; cognitive competence; moral competence; and spiritual competence. The 30-item CRS was considered appropriate as a newly developed instrument.


Introduction
As the caregivers' role, some caregivers of older persons with dementia have to provide care for 24 h per day and the long term care continuing for 15 years or more (Landefeld, Palmer, Johnson, Johnston, & Lyon, 2004). When having to live with long term care giving caregivers are likely to

PUBLIC INTEREST STATEMENT
Caregivers are the heart of care. Caring for older persons with dementia is associated with poor quality of life (QOL), poor physical and psychological health of caregivers. Resilience enabled caregivers to return to normal after traumatic psychological events, after change, after being exposed to serious events, and after a life threatening situation, making caregivers able to better cope with the demand of caring for elderly persons with dementia.
Health care professionals could also employ the Caregiver Resilience Scale as a guideline when assessing resilience among caregivers of older persons with dementia. Resilience outcomes are balance of mind and effective coping. Finally, they were able to continue providing care to older persons with dementia effectively. develop stress. Furthermore, a high level of a patient's dependence impacts both their physiological and psychological well-being (Adams & Smyth, 2008;Pinquart & Sorensen, 2007). A review of previous research associated with caregivers' burden of caregivers of older persons with dementia has revealed that resilience is one of the key buffers and protective factors for caregivers burnout and, strain as it, decreases caregivers burden or caregivers distress.
Resilience is a process of growth and adaption, and it has a multi-dimensional structure. It is also a holistic, dynamic, and development that encompass the ability to cope with stress, and, serious situation (e.g. family problems, family relationships, and serious illnesses). Resilience outcomes are a balance of mind and effective coping. In other words resilience causes emotional strength, especially psychological strength of caregivers during interactions with others who have distress, psychological traumas, or emotional uncertainty. Some caregivers have more resilience, strength, and resistance to stress than others caregivers. Theses caregivers have good coping ability and can develop strength from the stress and trouble they are facing. In addition, caregivers who have resilience have the ability to seek meaning, goals, and values during complicated situations (Siebert, 2007c). Assessment of individual resilience is commonly undertaken by use of a resilience scale. A review of previous research has shown that there are two instruments that are popularly used to assess individual resilience as follows: (a) the 25-item Connor-Davidson Resilience Scale (CD-RISC) (Connor & Davidson, 2003), and (b) the 22-item Response to Stressful Experiences Scale (RSES) (Johnson et al., 2011).
At the same time, the literature review has revealed limited use of the CD-RISC to assess resilience in the general population and the Stressful Experiences Scale (RSES) has been mostly used to assess resilience in the military.
It is worth noting that most studies on caregivers' resilience have been conducted in western countries and at present no research has been done on an assessment of caregiver resilience using a resilience scale with Thai caregivers of older persons with dementia. Correspondingly, the development of a resilience scale for caregivers of older persons with dementia is much needed.

Purpose
(1) To develop the caregiver resilience scale for Thai caregivers of older persons with dementia.
(2) To examine the validity and reliability of the caregiver resilience scale.

Exploring the resilience concept and specifying the domains of caregivers' resilience
The pre-specific domain of the Caregiver Resilience Scale (CRS) was determined by using the method of concept analysis from the literature review and exploration of caregivers' resilience (Davis, 1999;Grotberg, 1999;Gunnestad, 2006;Lester, Masten, & McEwen, 2006;Reivich & Shatte, 2002;Roland, 2006;Siebert, 2007aSiebert, , 2007b. In the end, the resilience theory offered by Davis (1999) was used as a guideline in this study. Then, a qualitative approach was conducted by semi-structured interview among 10 caregivers in order to confirm the pre-specified structure of caregivers' resilience and congruence with Thai caregivers' context.

Formatting the CRS
Based on the data from an exploration of the concept of caregivers' resilience and the Resilience Theory of Davis (1999), the first draft of the Caregivers Resilience Scale (CRS) developed in the present study was composed of six domains and 36 items, which were as follows: physical competence (items 1-6), relationship competence (items 7-12), emotional competence (items 13-18), moral competence (items 19-25), cognitive competence (items 26-31), and spiritual competence (items 32-36). These 36 items which were arranged in a four point Likert scale, with the response choices ranging from 0 (not true) to 3 (mostly true). The scoring criteria for positive items were described as follows: 0 (not true), 1 (sometimes true), 2 (frequently true), and 3 (mostly true), while the scoring criteria for negative items were as follows: 0 (mostly true), 1 (frequently true), 2 (sometimes true), and 3 (not true). Possible total scores ranged from 0 to 108 points, with higher scores indicating higher levels of caregivers' resilience.

The psychometric evaluation was adopted to evaluate the validity and reliability of the CRS
The content validity was validated by three professionals who gave expert opinions related to internal consistency with definitions of terms. The contents validity evaluation from which were created for the experts to give their opinion on the relevancy and clarity of each item which was a three point rating scale (−1 = not relevant, 0 = somewhat relevant, and 1 = quite relevant) was used.

Pre-test
The CRS was also tried out with 30 caregivers to test its internal consistency. Cronbach's alpha coefficient was used to estimate the internal consistency of the CRS. An alpha coefficient of 0.7 (Nunnally & Bernstein, 1994) or higher was considered acceptable for the newly developed instrument.

Field test
Item correlation was analyzed in the second round of evaluation of factor analysis. The items with a correlation of 0.3 were considered acceptable (Nunnally & Bernstein, 1994).
Exploratory Factor Analysis (EFA) was used to examine the number of domains and indicators of the CRS item set. Principle components and varimax methods were conducted to extract and rotate the components.
The rotate factors were interpreted by testing the items load at 0.30. The five criteria employed in evaluating items were eigenvalues, scree plot, and percentage of variance, factor loading, and theoretical interpretability. Therefore, the scores from the participants were considered to extract and specify factors of the CRS. Finally, the final draft of the Caregivers Resilience Scale (CRS) consisted of 30 items which could be categorized into six factors.

Participants and setting
The sample size was calculated based on the sample criteria of factor analysis. Munro (2005) suggested that the ratio of about five to ten subjects per item was adequate for an initial psychometric evaluation of a newly developed instrument. Finally, there were 150 participants.
The participants were 150 primary caregivers of older persons diagnosed with mild to moderate dementia or any subtypes of dementia. These patients all had appointments at the Memory Clinic, Neurological Clinic, or Geriatric Clinic in the Out Patient Department (OPD) at a regional hospital in upper southern Thailand. The participants were selected by means of purposive sampling according to the following inclusion criteria: (1) they were primary caregivers who provided the majority of care for older persons with dementia, (2) they were between aged 20 to 60 years of age, (3) they were able to speak and understand the Thai language, and (4) they agreed to participate in this study.

Protection of human subjects
Protection of human subjects was sought from the Institutional Review Board (IRB), the Ethics Review Committee for Research Involving Human Research Subjects, Health Science Group, Chulalongkorn University (COA No. 063/2557), and the Ethics Committee of the two Regional Hospital of upper southern Thailand.
The first Informed Consent Form (ICF) was used in the pre-test and field test to ensure the participants' willingness to participate in the semi-structure interview and psychometric evaluation.

Results
The first draft of the CRS was composed of 36 items within the six domains of (physical competence, relationship competence, emotional competence, cognitive competence, moral competence, and spiritual competence). The draft was examined by three experts who recommended the deletion of six items, which were considered redundant. After omitting the six items, the content validity index was equal to 0.84, which was acceptable relevance for the construct of caregivers' competence.
In the second drafts of the CRS, a total of 30 items was retained. The draft was also tried out with 30 caregivers to test its internal consistency, and the result showed that the internal consistency was at a high level (Cronbach's alpha 0.87). For the item analysis: Cronbach's alpha coefficient was employed to examine the correlation between item and item, item and subscale, and item and total. Item correlations ranging from 0.56 to 0.88 were considered acceptable. Therefore, the correlation between item and total scores was accepted as adequate.
The third draft of the 30-items CRS was examined to ensure internal consistency. The alpha of each subscale and the total scale were evaluated. An alpha coefficient of the total CRS was 0.87, while the alpha coefficients of the subscales of the physical competence domain, relationship competence domain, emotional competence domain, moral competence domain, cognitive competence domain, and spiritual competence domain ranged from 0.52 to 0.87. The correlations of items and their subscale ranged from 0.32 to 0.83.
The results of Bartlett's test of sphericity showed a significant high inter-item correlation (χ 2 = 17,124.13, p < 0.01). The KMO reported the sampling adequate at 0.91 which was an excellent indicator of factorability. Factor extraction: In this step, the communalities of the 30 items of factor extraction ranged from 0.40 to 0.78. The six components with initial eigenvalues greater than 1 ranged from 1.43 to 21.34, and the total variances at 63.67 were reported.
The rotation using the orthogonal varimax method was calculated in this step. The result was considered a low level in the hierarchy of factors. The low level factor seemed less meaningful to the overall analysis (Comrey & Lee, 1992). Factor rotation was used to improve interpretation. After rotation, the six components showed a total percentage variance of 63.67% as presented (Table 1). The items statements and communality of each item in the CRS with all factors are provided (Table 2).

Discussion
The final version of the CRS for Thai caregivers was composed of 30 items within six domains: physical competence; relationship competence; emotional competence; cognitive competence; moral competence; and spiritual competence. The existing theoretical structures of the CRS are described as below: Domain I: Physical competence was the sixth contributing factor to caregivers resilience in Thai primary caregivers (% of variance = 13.10). The study findings indicated that Thai primary caregivers did not perform self-care by using physical competence. This was simply because Thai primary caregivers were mainly blood relatives of the patients, so they tended to devote their time for the care of their loved one. Therefore, healthcare professionals should promote the physical health of caregivers and encourage them to perform self-care so as to maintain their health status (Ritteveerakul, 2005).
Domain II: Relationship competence was the fifth-ranked contributing factor to resilience (% of variance = 15.60). All items reflected communication skills. From the literature review, the point was made that with older persons with dementia communication is important because it has effects on the outcome of caring for older persons with dementia (Savundranayagam, Hummert, & Montgomery, 2005).
Domain III: Emotional competence was the third contributing factor to resilience (% of variance = 18.31). The findings indicated that Thai primary caregivers had to bear the direct effect of caring for older persons with dementia. Similarly, previous research (Sörensen, Duberstein, Gill, & Pinquart, 2006) has reported that the direct strain from caring for older persons with dementia on health status includes depression, anxiety, stress, headache, sleeplessness, and exhaustion.
Domain IV: Moral competence was the second contributing factor to resilience in Thai primary caregivers (% of variance = 22.31). All items in this domain reflected caring for older persons with dementia. According to the Resilience Theory of Davis (1999) and a literature reviews, moral competence is important because moral competence enables caregivers to provide complete care that can control or delay progression of the disease.
Domain V: Cognitive competence was the fourth factor to resilience in Thai primary caregivers (% of variance = 15.70). It is worth noting that behavioral and psychological management was frequently found in mild to moderate demented elderly persons. Therefore, cognitive competences is important for primary caregivers as it equips them with the' ability to properly carry out behavioral and psychological management.

Item
Item statement Communality